The overall goal of this project is to determine predictors of traffic-related air pollution inside and outside of the homes of urban residents, to reduce exposure misclassification for assessments of asthma etiology as well as other intraurban epidemiological analyses. For multifactorial diseases like asthma, it is impossible to determine potential multiplicative or synergistic effects without refined exposure data. Conventional exposure paradigms in this context either rely on proximity to roadways, which is not pollutant specific, or on measurements from central site monitors, which do not capture small-scale exposure heterogeneity. For this study, air pollution measures have been taken among a subset of participants in an asthma birth cohort study, with sampling sites selected using GIS-based traffic characterization. Along with indoor and outdoor measurements of nitrogen dioxide, elemental carbon, fine particulate matter, and polycyclic aromatic hydrocarbons, information has been collected on home characteristics and occupant behaviors. We hypothesize that both ambient monitoring data and traffic characteristics will predict outdoor levels of these air pollutants, using statistical models that address both spatial and temporal variability. In addition, we hypothesize that indoor levels of these pollutants will be related to both levels outside of the home and home/occupant characteristics, but that use of only publicly-available GIS data will not significantly reduce measurement error. This modeling approach will demonstrate what information must be collected by investigators to accurately assess air pollution exposures in urban settings. This dataset provides a unique opportunity to understand exposure patterns in urban settings, given 1 of the largest multi-pollutant samples to date that includes both indoor and outdoor measures, which can inform the design and implementation of future epidemiological investigations. [unreadable] [unreadable]